In recent years, an autonomous distributed network, such as an ad hoc network, has been used, in which a network is established by directly connecting terminals to one another without using network infrastructures, such as base stations or access points.
The autonomous distributed network connects a large number of terminals without using access points, based on a technology applied to wireless connections between computers, personal digital assistants (PDAs), mobile telephone terminals, and the like. The autonomous distributed network autonomously establishes a network due to addition or deletion of nodes. Therefore, the autonomous distributed network can reduce load for setting a network and is useful when a large number of nodes are provided or when a network configuration is changed with some frequency.
As a technology for managing the network status of the autonomous distributed network as described above, there is a known method to cause a management server to centrally manage the network status or cause an individual node to autonomously manage the network status by collecting neighborhood topologies.
Patent Literature 1: International Publication Pamphlet No. WO 2005/117365
Patent Literature 2: Japanese Laid-open Patent Publication No. 2003-061120
However, in the conventional technology, the management server or each of the nodes transmits a packet to other nodes every time the management server or each of the nodes detects an event, such as initial start-up of a node, addition of a node, or a network disconnection. Therefore, there is a problem in that a traffic increases and load on the nodes increases.
For example, when detecting an event, the management server or each of the nodes transmits a packet to all of the nodes in a network and receives responses. Subsequently, the management server or each of the nodes generates a topology based on the received responses. Thereafter, the management server or each of the nodes compares a topology that is generated by detection of a previous event with the topology that is generated at this time to detect a change in the network configuration.
In this way, the management server or each of the nodes performs a large number of processes to specify a change in the network configuration. Therefore, in the large-scale network including an enormous number of nodes, the processing load increases enormously. Furthermore, in the large-scale network, an enormous number of nodes transmit packets, so that the traffic volume increases enormously. Moreover, if nodes are added or deleted frequently, there is a problem in that the processing load becomes a bottleneck and it becomes impossible to follow a change in the network configuration.